1. Field
Exemplary embodiments relate to techniques of encoding a screen, and transmitting the encoded screen to a terminal in a server constituting a cloud system. More particularly, exemplary embodiments relate to techniques of analyzing a region of a window displayed on a screen by using application programming interface (API) information processed by a kernel or Operating System (OS), and also encoding the screen by using depth information of the window.
2. Description of the Related Art
Data encoding methods of the related art may be classified into lossy encoding methods, and lossless encoding methods. When a region, such as a text, an icon, etc., having a feature in which even a slight deterioration in image quality is conspicuous is compressed, lossless encoding methods are generally preferred. When a region, such as a photograph, a moving picture, etc., in which a pixel structure is complicated is compressed, lossy encoding methods are preferred because a data loss is not conspicuous, even though data may be lost.
However, when a compound image, including both a graphic element (a text, an icon, a figure, etc.) and an image element (a photograph, a moving picture, etc.), is encoded in the related art, the compound image is usually encoded using one algorithm. For example, a moving picture or photographic image encoding algorithm, such as H.264, MPEG-2, or JPEG, or a graphic element suitable encoding algorithm may be uniformly applied. In addition, when individual regions of a compound image are selectively encoded, the compound image may be analyzed, and encoding algorithms suitable for regions may be selectively applied according to the analysis result. In this case, an image analyzing process is mandatory. For image analysis, low-level image processing is representatively used. A low-level image processing technique of the related art analyzes a screen using statistical characteristic information of pixels and planarization or a gradient level of an image, which are obtained by using pixel values acquired by scanning a region to be encoded.
In a related art cloud system, a control signal input from a client terminal is processed by a server, and a processing result is displayed on a screen of the client terminal. In many cases, the client terminal does not have a separate Central Processing Unit (CPU) or a rendering tool. In this case, the server encodes an image to be displayed, and transmits the encoded image to the client terminal. However, a cloud display screen is usually provided with a compound image, including a text, a graphic, an image, a moving picture, etc.
If an encoding method of the related art is used for the cloud display screen, regardless of image characteristics, the encoding method may not be suitable. For example, it is assumed that a documentation work is performed using a cloud terminal. When an image analysis is applied to a cloud display screen on which the documentation work is being performed, text has a high probability to be analyzed as a text region, because the text has a large pixel value difference. However, if a related art image encoding algorithm is used, image quality may be significantly deteriorated, and readability of the text may decrease to a level at which it is difficult to perform the documentation work. Thus, it is very important to apply an encoding algorithm suitable for each region by using an image analysis. However, the low-level image processing technique of the related art always has a false alarm probability, and a missing alarm probability in region identification.